ZYH-LLM-Qwen2.5-14B-V3

This is the third-generation model of the ZYH-LLM series.

It employs a large amount of model merging techniques, aiming to provide a powerful and unified 14-billion-parameter model, laying a solid foundation for further model merging and model fine-tuning.

As of February 25, 2025, the 14B model with the highest IFEval score

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	41.63
IFEval (0-Shot)	85.78
BBH (3-Shot)	48.18
MATH Lvl 5 (4-Shot)	52.72
GPQA (0-shot)	10.96
MuSR (0-shot)	9.00
MMLU-PRO (5-shot)	43.12

The following are the specific details of model merging, hoping to inspire you:

First stage:

Step 1:

models:  
  - model: Qwen/Qwen2.5-14B-Instruct  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: Qwen/Qwen2.5-14B  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: Qwen2.5-14B-YOYO-1010

models:  
  - model: Qwen/Qwen2.5-14B-Instruct-1M  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: Qwen/Qwen2.5-14B  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: Qwen2.5-14B-YOYO-1010-1M

models:  
  - model: Qwen/Qwen2.5-14B-Instruct  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: EVA-Qwen2.5-14B-YOYO-1010

models:  
  - model: Qwen/Qwen2.5-14B-Instruct-1M  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: EVA-Qwen2.5-14B-YOYO-1010-1M

Step 2:

models:  
  - model: EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: Qwen/Qwen2.5-14B  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: EVA-Qwen2.5-14B-base

merge_method: sce  
models:  
  - model: EVA-Qwen2.5-14B-base  
base_model: Qwen/Qwen2.5-14B-Instruct-1M  
parameters:  
  select_topk: 1  
dtype: bfloat16  
tokenizer_source: base  
normalize: true  
int8_mask: true  
name: Qwen2.5-14B-pro

Step 3:

models:
  - model: Qwen2.5-14B-YOYO-1010-1M
  - model: Qwen2.5-14B-YOYO-1010
  - model: EVA-Qwen2.5-14B-YOYO-1010-1M
  - model: EVA-Qwen2.5-14B-YOYO-1010
merge_method: sce
base_model: Qwen2.5-14B-pro
parameters:
  normalize: true
  int8_mask: true
dtype: bfloat16
tokenizer_source: base
name: ZYH-LLM-Qwen2.5-14B-V3-preview

Second stage:

models:  
  - model: Qwen/Qwen2.5-14B-Instruct  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: arcee-ai/Virtuoso-Small-v2  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: Qwen2.5-14B-YOYO-della1

models:  
  - model: Qwen/Qwen2.5-14B-Instruct-1M  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: arcee-ai/Virtuoso-Small-v2  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: Qwen2.5-14B-YOYO-della2

models:  
  - model: Qwen/Qwen2.5-14B-Instruct  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: Azure99/Blossom-V6-14B  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: Qwen2.5-14B-YOYO-della3

models:  
  - model: Qwen/Qwen2.5-14B-Instruct-1M  
    parameters:  
      density: 1  
      weight: 1  
      lambda: 0.9  
merge_method: della  
base_model: Azure99/Blossom-V6-14B  
parameters:  
  density: 1  
  weight: 1  
  lambda: 0.9  
  normalize: true  
  int8_mask: true  
dtype: bfloat16  
tokenizer_source: base  
name: Qwen2.5-14B-YOYO-della4

Final stage:

merge_method: model_stock
base_model: ZYH-LLM-Qwen2.5-14B-V3-preview
models:
  - model: Qwen2.5-14B-YOYO-della1
  - model: Qwen2.5-14B-YOYO-della2
  - model: Qwen2.5-14B-YOYO-della3
  - model: Qwen2.5-14B-YOYO-della4
dtype: bfloat16
tokenizer_source: base
int8_mask: true
normalize: true
name: ZYH-LLM-Qwen2.5-14B-V3